Pytorch 神经网络错误

Question

I am trying to compute the loss and accuracy of a certain machine learning model by using Pytorch and I am having trouble initializing the dataset so that it can run. Using the Moon dataset, I am getting a few errors when I run the code. I first initialize the dataset:

X, y = make_moons(200, noise=0.2, random_state=42)

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.30, random_state=1, stratify = y)
x, y = Variable (torch.from_numpy(X_train)).float(), Variable(torch.from_numpy(y_train)).float()

and then when I run the Neural Network:

    def __init__(self):
        super(SoftmaxRegression, self).__init__()
        self.fc = nn.Linear(200, 1)
        self.softmax = nn.Softmax()

    def forward(self, x):
        x = self.fc(x)
        x = self.softmax(x)
        return x

I get the following errors: serWarning: Implicit dimension choice for softmax has been deprecated. Change the call to include dim=X as an argument. x = F.softmax(self.layer(x))
ret = torch._C._nn.nll_loss(input, target, weight, _Reduction.get_enum(reduction), ignore_index) IndexError: Target 1 is out of bounds.

How can I fix this so that it can run the dataset and output the loss and accuracy?

Answer 1

(Sorry to put this as an answer but unfortunately stack overflow won't let me comment:/).

Even if the Softmax worked it is absolutely pointless.(Unless you are softmaxing across your batch but that would be really weird). Your code shows you have a linear layer going from a tensor of 200 to 1. Softmax on a single value will simply return that value, Softmaxing should only be used on 2 or more values.

If you wish to do binary classification I would instead change the code to be this:

    import torch.nn.functional as F

    def forward(self, x):
        x = self.fc(x)
        x = F.sigmoid(x)
        return x